[vim/vim] Unicode EN (half an EM) support (#3775)
Nobuhiro Takasaki
In Unicode, there are only three characters whose display width should be one cell. (EN QUAD / EN SPACE / EN DASH).
It is a problem with CJKV.
ambiwidth should not be applied to these characters.
When it is applied, it looks like this.
This screen is written in Japanese.
(EN DASH) is used. It is a short horizontal bar.
In this example, ambiwidth is 2.
In ConPTY at Vim, the rows are swapped.
In winpty in Vim, the rightmost character is not output.
In cmd.exe, it will display normally.
When this patch is applied, the cell width of (EN DASH) becomes 1 and it is displayed normally.
You can view, comment on, or merge this pull request online at:
https://github.com/vim/vim/pull/3775
Commit Summary
- Unicode EN (half an EM) support
File Changes
- M src/mbyte.c (11)
Patch Links:
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
Codecov
Codecov Report
Merging #3775 into master will increase coverage by
<.01%.
The diff coverage is100%.
@@ Coverage Diff @@ ## master #3775 +/- ## ========================================== + Coverage 78.37% 78.37% +<.01% ========================================== Files 102 102 Lines 141275 141277 +2 ========================================== + Hits 110726 110731 +5 + Misses 30549 30546 -3
| Impacted Files | Coverage Δ | |
|---|---|---|
| src/mbyte.c | 62.97% <100%> (+0.04%) |
⬆️ |
| src/if_xcmdsrv.c | 84.17% <0%> (-0.18%) |
⬇️ |
| src/ex_cmds2.c | 84.49% <0%> (-0.1%) |
⬇️ |
| src/message.c | 75.94% <0%> (-0.05%) |
⬇️ |
| src/window.c | 83.39% <0%> (+0.03%) |
⬆️ |
| src/ui.c | 47.37% <0%> (+0.07%) |
⬆️ |
| src/terminal.c | 74.44% <0%> (+0.08%) |
⬆️ |
| src/gui.c | 57.98% <0%> (+0.15%) |
⬆️ |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact),ø = not affected,? = missing data
Powered by Codecov. Last update 04b4e1a...9c43f92. Read the comment docs.
Nobuhiro Takasaki
(EN QUAD) is C/D normalized (EN SPACE).
(EN SPACE) is KC/KD normalized to U+0020.
Both are U+0020, which is a blank space of cell width 1.
Nobuhiro Takasaki
After applying the patch.
(EN DASH) with a red underline.
mattn
@mattn commented on this pull request.
In src/mbyte.c:
> @@ -1414,6 +1414,15 @@ utf_uint2cells(UINT32_T c)
int
utf_char2cells(int c)
{
+ /* Sorted list of non-overlapping intervals of EN(single width characters
+ * in typesetting) that are not affected by p_ambw and CJKV. */
+ static struct interval singlewidth[] =
+ {
+ {0x2000, 0x2000},
+ {0x2002, 0x2002},
+ {0x2013, 0x2013}
+ };
I don't understand why you make this list. Just removing the entries from variable ambiguous? Also I noticed \u2013 is defined as A in http://www.unicode.org/Public/11.0.0/ucd/EastAsianWidth.txt
mattn
@mattn commented on this pull request.
In src/mbyte.c:
> @@ -1414,6 +1414,15 @@ utf_uint2cells(UINT32_T c)
int
utf_char2cells(int c)
{
+ /* Sorted list of non-overlapping intervals of EN(single width characters
+ * in typesetting) that are not affected by p_ambw and CJKV. */
+ static struct interval singlewidth[] =
+ {
+ {0x2000, 0x2000},
+ {0x2002, 0x2002},
+ {0x2013, 0x2013}
+ };
\u2000 and \u2002 both not exists in doublewidth/emoji_width/ambiguous.
Nobuhiro Takasaki
@ntak commented on this pull request.
In src/mbyte.c:
> @@ -1414,6 +1414,15 @@ utf_uint2cells(UINT32_T c)
int
utf_char2cells(int c)
{
+ /* Sorted list of non-overlapping intervals of EN(single width characters
+ * in typesetting) that are not affected by p_ambw and CJKV. */
+ static struct interval singlewidth[] =
+ {
+ {0x2000, 0x2000},
+ {0x2002, 0x2002},
+ {0x2013, 0x2013}
+ };
I listed the letters with EN attached. It was a coarse patch. I will review again.
Nobuhiro Takasaki
\u2013 for special processing. Otherwise, it can not be compatible with cmd.exe.
An indication from @mattn, mintty draw 2 cells as ambiguous width.
I will think further.
I thought about now that we can set the cell width of \u2013 in the variable 'enwidth'.
Or special processing by detecting cmd.exe.
Otherwise it will allow destruction of the display.
I feel like I should not do anything...
mattn
I wrote some ideas in this comment: #2074 (comment)
Nobuhiro Takasaki
I think that it is good design to read the text file which defined the range.
Since I write a demonstration sample, I want to advance the topic.
If this is not over, I will not be able to proceed with ConPTY because of the display.
Nobuhiro Takasaki
My act is to make a roof on the roof.
I do other things to do.
I need more thought.
Nobuhiro Takasaki
@ntak pushed 2 commits.
—
You are receiving this because you are subscribed to this thread.
Nobuhiro Takasaki
I started with the detection of cmd.exe.
Cause of the screen being incorrect at the compatible terminal, now there is also a patch for it, this will be divided into different patches.
I keep ConEmu's detection in, so I will be a little slovenly, I will be careful.
Nobuhiro Takasaki
I think that it was able to correspond to the whole query.
It would be cool if this was available from the beginning.
The width of EN DASH is Japanese-specific. The value of this patch itself is questioned.
EN DASH becomes one cell only when cmd.exe is Japanese. There are infinite combinations.
It is useful if you can use queries, think about other uses.
Please ignore this issue.
At a later date, reapply as a patch with another purpose and link.
Christian Brabandt
so closing.
Christian Brabandt
Closed #3775.
mattn
Ref: 08aac3c
—
Reply to this email directly, view it on GitHub.
You are receiving this because you are subscribed to this thread.